大规模时变区域覆盖优化建模及其高性能求解

Regional coverage models are widely used in emergency management, business siting, and public transportation planning. Under dynamic conditions,the modeling of regional coverage problems is much more complicated than its static counterparts.Meanwhile in reality, the large numbers of service facilities and demands, plus their dynamic distribution in geographic space and time-varying uncertainty,optimization models are usually complex. Therefore it is difficult to solve these models efficiently and to meet the practical needs. In this proposal, we aim to tackle the modeling of dynamic time-dependent coverage problems and the solving of large-scale complex coverage models. Based on our previous research experience in spatial optimization, high performance gocomputation, and traffic data analysis, we strive to build a comprehensive regional coverage modeling framework that is capable of handle dynamic traffic conditions. The reduction and efficent solving of large scale complex coverage models will also be investigated. Further, we will focus on the development of parallel algorithms for both exact and heursitic solutions. Finally, emergency medical service location and public transit planning improvement are selected as case studies to demonstrate the proposed methodology. Our research would underpin various regional coverage applications by providing efficient and applicable methods and approaches. Additionally, our research findings will significantly advance spatial optimization and high-performance geocomputation theories and techniques.

区域覆盖优化模型是应急资源分配、商业选址和公共交通规划等空间决策应用中常见的空间优化模型。动态交通条件下，区域覆盖优化问题建模远较传统静态模型复杂。同时大规模实际应用中，服务设施以及需求数量巨大，时空分布多样，还需考虑时变不确定性情况，造成优化模型形式复杂，求解困难，难以满足快速优化求解的实际需要。本研究针对动态时变覆盖优化建模以及大规模复杂模型求解两个问题，基于项目组在空间优化、高性能地理计算和交通时空数据分析的有关研究基础，建立顾及交通状态的时变区域覆盖优化模型框架，提出大规模复杂模型的化简与高效求解流程方法，研究精确解和启发式近似解的高性能并行求解方法和策略。本研究拟采用应急医疗服务和公共交通优化改善两个应用问题，通过实际数据验证提出的理论与方法。研究成果将为区域覆盖优化的各种应用提供高效可行的方法与技术支持，并有力地推动空间优化决策和高性能地理计算理论方法的发展。

区域覆盖优化是一种常见的空间优化决策技术，可广泛应用于应急资源分配、商业选址和公共交通规划等领域。但是传统区域覆盖优化模型没有充分考虑动态交通条件的覆盖范围变化情况，对于大规模城市复杂模型的建模和求解也缺乏良好的解决方法。本项目针对大规模区域覆盖优化模型在动态交通条件下的建模和求解问题，建立了面向大城市复杂道路网络与实际交通情况的时变区域覆盖优化高可扩展模型框架，提出了面向动态交通条件的公交可达指标，设计实现了时变可达量度为核心的实用化时变区域覆盖优化模型，提出了顾及隐含的时空信息的模型启发式求解方法。采集了武汉和深圳市的基础地理信息和浮动车数据，探索和验证了大规模城市覆盖优化问题的高性能求解算法。针对时变不确定交通条件，结合应急医疗和公交规划改善问题验证了以上提出的模型和算法。项目组经过四年的研究，基本完成了原定的研究目标，掌握了实际数据驱动的覆盖优化建模、求解和分析策略方法，初步建立起一套复杂区域覆盖优化模型的理论与技术方法体系。此外还在项目支持下研究了公交服务的公平量度方法、路网特征与覆盖优化之间的关系以及基于新型高性能计算平台的地理计算技术。研究成果不仅在理论和方法上有力地推动了空间优化决策和高性能地理计算的发展, 还可为区域覆盖优化的实际应用提供高效可行的方法与技术支持。

内容获取失败，请点击重试